The Rise Of The Machines?

I recently had lunch with an old friend.  We do that sort of thing aperiodically; but when we get together the event is “scripted”.  We meet at the same restaurant; we tell the server to take her/his time because we are going to be there for a while.  We order an adult beverage, ask perfunctory questions about the families and then spend the next two hours talking sports.  That sequence is about as reliable as death, taxes and Stephan A. Smith getting outraged over something/anything at around 10:00 AM ET every weekday.

We were lamenting the negative effect that analytics has had on baseball – – actually we were lamenting the slavish addiction to analytics that is hurting baseball.  It seems that more and more managers fail to recognize that the game taking place in front of them on a given day is not an average game nor one that necessarily fits a trend.  What used to be called “over-managing” when Alvin Dark and Gene Mauch were in opposite dugouts has now become too much sway given to the “numbers nerds”.

And then the conversation took a dark turn.  We wondered how long it will be until Artificial Intelligence wiggles its way into sports.  I certainly do not want that day to happen any time soon, but I must admit that having AI fiddle with sporting events is a lot less threatening than having AI fiddle with the nuclear launch codes.  We sort of agreed that AI would probably not find as comfortable a perch in sports like basketball or ice hockey or soccer but that there might be niches for it to occupy in baseball and in football.  Since our lunch was proximal to the last NFL Draft, my friend asked if AI might not become a contributor for some teams as they construct their draft boards.

I think he has hit upon a role for AI in an important NFL event because the Draft is indeed an important way to build a roster – – and – – blown draft picks and derail a “team on the rise” rather quickly.  When you consider that half of the first-round picks in a typical NFL Draft will not pan out nearly to the extent expected, it would seem as if any sort of a “boost” from outside would be welcomed and adopted.

Let me get a disclaimer in here quickly:

• Neither my friend nor I would qualify even as naïfs when it comes to AI.

If indeed Artificial Intelligence can learn on its own beyond the coded rules that set up the foundational machinery, then maybe the large computing power of an AI system can digest game film from college players and more accurately project those skills and abilities to a benchmark NFL level.  I am not talking here about a computer’s ability to measure things more accurately than a human can; I don’t think it matters all that much if Player A has a reaction time that is a hundredth of a second faster than Player B.  But perhaps an AI system might be able to assimilate a far greater number of observable variables and make connections that have gone unrecognized to date.

Even more “radical” might be an AI system that looked at the observables that make for successful NFL players.  Call it an idealized player sort of like a reference electrode in electrochemistry.  If a team were to think it had a handle on what sorts of things it took physically, mentally and in terms of character to become a successful NFL player, it could then have a template by which it might evaluate potential draftees.

Before you tell us this is all pie in the sky, remember what the current modus operandi produces:

• Half of the FIRST-ROUND picks will be mediocre at best in the NFL.

The standard that an AI system has to meet or beat is not all that stringent.

Moving on …  There will be two new rules in effect for NFL games this year.  Hopefully, one of them will never need to come into play and the other of them will not come into play universally.  Call the first one the “2022 NFC Championship Game Rule”.  Recall that the Niners lost both of their QBs to injury in that game and basically played out the string without a QB who had college experience at the position.  That situation is not difficult to fix, and the owners have installed a new rule to take care of it.  Actually, what the owners did was to go back to an older rule:

• Up until 2012, teams would dress 45 players for a game and an “emergency QB” who could only enter the game in prescribed ways.
• Then the NFL just expanded the roster number to 46 and let teams decide if they wanted an “emergency QB” or if they preferred an “extra DB”.
• Now, the teams will be allowed to dress 46 players PLUS an “emergency QB”.

Presumably, that is a solved problem…

The other rule change should drastically reduce the number of kickoff return attempts.  At present, a kickoff that goes out of the end zone is put in play at the 25 yardline.  Under a new rule this year, any returner can call for a fair catch at any point on the field behind the 25 yardline and the ball will still be put in play at the 25 yardline.  Seemingly, the objective here is minimize the number of kickoff returns which are considered to be less safe than ordinary plays from scrimmage.

I hope this does not remove the kickoff return from the game entirely.  This is the rule in college football and the number of return attempts at that level is significantly smaller than in the NFL.  I am not sure if I like this change or not …

Finally, having opined today on AI, let me close with this from F. Scott Fitzgerald on how to measure real intelligence:

“The test of a first-rate intelligence is the ability to hold two opposed ideas in the mind at the same time, and still retain the ability to function.”

But don’t get me wrong, I love sports………